View
3
Download
0
Category
Preview:
Citation preview
1 | P a g e
Investigating the impact of flow rate ramp-up on Carbon Dioxide start-up injection Revelation J Samuel1 and Haroun Mahgerefteh
Department of Chemical Engineering, University College London, WC1E 7JE, UK
Received / Accepted / Published
Abstract: Carbon Capture and Storage (CCS) represents the technology for capturing carbon dioxide
(CO2) produced from large emission sources, such as fossil-fuel power plants, transporting and
depositing it in underground geological formations, such as depleted oil and gas fields. CO2 injection
flow rate ramp-up time is essential for the development of optimal injection strategies and best-
practice guidelines for the minimisation of the risks associated with the process. The rate of rapid
quasi-adiabatic Joule-Thomson expansion when high pressure CO2 is injected into a low pressure
injection well if not monitored carefully may lead to significant temperature drops posing several
risks, including: blockage due to hydrate and ice formation with interstitial water. The paper employ
a Homogeneous Equilibrium Mixture (HEM) model, where the mass, momentum, and energy
conservation equations are considered for a mixture of liquid and gaseous phases assumed to be at
thermal and mechanical equilibrium with one another. In particular, this study considers linearly
ramped-up injection mass flow rate from 0 to 38.5 kg/s over 5 minutes (fast), 30 minutes (medium)
and 2 hours (slow). The reliability and applicability of the HEM model is tested against real-life CO2
injection experiment wellbore temperature data obtained from the Ketzin pilot site Brandenburg,
Germany. The ramping up injection mass flow rate simulation results predicted the fast (5 mins)
injection ramp-up as best option for the minimisation of the associated process risks.
Keywords: boundary conditions; carbon capture and storage; CO2 injection; ramp-up; transient
model
1 This work was supported in part by the Petroleum Technology Development Fund (PTDF) under Grant PTDF/E/OSS/PHD/SRJ/717/14 and the Department of Chemical Engineering, University College London. Corresponding author: Revelation Samuel specialises in multi-phase flow modelling, pipeline safety and risk management and CCS. Email: ucecrjs@ucl.ac.uk; Tel.: +44 2076793809
2 | P a g e
1. Introduction
Carbon Capture and Storage (CCS)
represents the technology for capturing carbon
dioxide (CO2) produced from large point
sources, such as fossil-fuel power plants,
transporting it, e.g. by pipelines, ships or
trucks, and depositing it in underground
geological formations, such as depleted oil and
gas fields or saline aquifers. CCS is widely
recognised as a key technology to meet the
ambitious 2 ºC agreement reached at the Paris
COP 21 meeting (COP 21, 2015; UNFCCC,
2016). While CCS technology is promising, it
is also very expensive, particularly due to the
high costs of capture and compression, which
have been the main focus of recent studies
(Dennis Gammer, 2016). In addition, it has
been pointed out that in order to reduce the
costs of CCS, the integration between the
various elements of its chain should be
considered (IEAGHG, 2015). While this has
been done mainly for CO2 capture/purification
and transport processes, less attention has been
paid to interfacing between the transportation
and storage elements of CCS, which has
become particularly important for the safe and
optimal injection of CO2 into depleted gas
fields (De Koeijer et al, 2014).
Depleted gas fields represent prime potential
targets for the large-scale storage of captured
CO2 emitted from industrial sources and fossil-
fuel power plants (Oldenburg et el, 2001). For
instance, in the case of the UK, the Southern
North Sea and the East Irish Sea depleted gas
reservoirs provide 3.8 of the total 4-billion-
tonne storage capacity required to meet UK’s
CO2 reduction commitments for the period
2020-2050 (Hughes, 2009). Depleted gas
reservoirs are often considered as preferential
for CO2 storage, given their proven capacity to
retain buoyant fluids and the availability of
geological data, such as pressure, porosity and
permeability, derived from years of gas
production (Sanchez Fernandez et al., 2016).
The UK has completed three large FEED study
projects for offshore CO2 storage at Hewett,
Goldeneye and Endurance sites (Cotton, Gray,
& Maas, 2017). In order to further the
development of CCS and boost the confidence
of both private investors and the public in the
deployment of full-chain industrial CCS
systems, it is of paramount importance to
guarantee that the third element of the chain,
i.e. the storage site, will be of high-quality and
operate in safe conditions. A recent study has
investigated the key factors having the highest
impact on the safe storage of CO2 into depleted
oil and gas fields (Hannis et al, 2017). In
particular, key geological storage properties,
such as pressure, temperature, depth and
permeability, can affect injectivity and lead to
variations in the CO2 flow. Therefore, changes
in the operation of the storage sites will require
flow and pressure management within the CO2
transportation network (Sanchez Fernandez et
al., 2016).
The most effective way of transporting the
captured CO2 for subsequent sequestration
using high-pressure pipelines is in the dense
phase (Böser & Belfroid, 2013). The
CO2 arriving at the injection well will typically
be at pressure greater than 70 bar and at
temperature between 4 and 8 ºC. Given the
substantially lower pressure at the wellhead,
the uncontrolled injection of CO2 will result in
its rapid, quasi-adiabatic Joule-Thomson
expansion leading to significant temperature
drops (Curtis M. Oldenburg, 2007). This
process could pose several risks, including:
blockage due to hydrate and ice
formation following contact of the cold
sub-zero CO2 with the interstitial water
around the wellbore and the formation
water in the perforations at the near
well zone;
possible freezing of annular fluids (see
Fig. 2) resulting in potential wellbore
damage.
thermal stress shocking of the wellbore
casing steel as it cools beyond its
specified lowest temperature leading to
its fracture and escape of CO2;
CO2 backflow into the injection system
due to the violent evapouration of the
expanding liquid CO2 upon entry into
the low pressure wellbore.
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
3 | P a g e
In order to minimise all these risks, accurate
mathematical models are necessary to simulate
the time-dependent behaviour of the CO2
injected into the well and subsequently released
into the depleted reservoir, and thereby develop
appropriate injection strategies to reduce the
subsequent risks.
A significant amount of research has already
been devoted to the analysis of the injection of
CO2 from wells into underground reservoirs
(André et al, 2007; Goodarzi et al, 2010;
Nordbotten et al, 2005). However, as noted in
Linga & Lund, (2016) and Munkejord et al,
(2016), the analysis of the transient behaviour
of the CO2 flowing inside the injection wells
has received limited attention in the literature,
and more experimental data are needed (De
Koeijer et al., 2014). Lu & Connell, (2008)
discussed a steady-state model to evaluate the
flow of CO2 and its mixtures in non-isothermal
wells. Paterson et al, 2008) simulated the
temperature and pressure profile considering
CO2 liquid-gas phase change in static wells,
assuming a quasi-steady state and neglecting
changes in kinetic energy. Lindeberg, (2011)
proposed a model combining Bernoulli’s
equation for the pressure drop along the well
and a simple heat transfer mechanism between
the fluid and the surrounding rock. The model
was applied to the Sleipner CO2 injection well.
Pan et al, (2011) presented analytical solutions
for steady-state, compressible two-phase flow
through a wellbore under isothermal conditions
using a drift flux model. Lu & Connell, (2014)
investigated a non-isothermal and unsteady
wellbore flow model for multispecies mixtures.
Ruan et al., (2013) and (Jiang et al., 2014)
developed a two-dimensional radial numerical
model to study CO2 temperature increase
mechanisms in the tubing, and the impact of
CO2 injection on the rock temperature in both
axial and radial directions. Xiaolu Li, Xu, Wei,
& Jiang, (2015) developed a model to account
for the dynamics of CO2 into injection wells
subject to highly-transient operations, such as
start-up and shut-in. Linga & Lund, (2016)
discussed a two-fluid model for vertical flow
applied to CO2 injection wells, predicting flow
regimes along the well and computing friction
and heat transfer accordingly. Xiaojiang Li et
al., (2017) presented a unified model for
wellbore flow and heat transfer in pure CO2
injection for geological sequestration, EOR and
fracturing operations. Also recently, Samuel &
Mahgerefteh, (2017) published a paper on
transient flow modelling of the start-up
injection process without ramp-up. The study
shows the consequent effect of quasi-adiabatic
Joule-Thomson expansion when high pressure
CO2 is injected into a low pressure injection
well. Acevedo & Chopra, (2017) conducted a
prior work on transients and start up injection
for the Goldeneye injection well. They studied
the influence of phase behaviour in the well
design of CO2 injectors. The study shows that
during transient operations (closing-in and re-
starting injection operations), a temperature
drop is observed at the top of the well for a
short period of time. This is due to the
reduction of friction caused by a lower
injection rate. The duration of these operations
dictates the extent to which the various well
elements are affected. The sequence of steady
state injection, closing-in operation (30
minutes), closed-in time (10minutes), and
starting-up operation (30minutes) was
simulated for the low reservoir pressure case
using OLGA. However, no publication
critically considered the flowrate ramping up
times studied in this paper.
In this study, we investigate the CO2
injection flow rate ramp-up times which is
driven by the need for the development of
optimal injection strategies and best-practice
guidelines for the minimisation of the risks
associated with the process. Investigating
injection flow rate ramp-up is essential in
understanding the rate of rapid, quasi-adiabatic
Joule-Thomson expansion when high pressure
CO2 is injected into a low pressure injection
well. A Homogeneous Equilibrium Model
(HEM), where the mass, momentum, and
energy conservation equations are considered
for a mixture of liquid and gaseous phases
assumed to be at thermal and mechanical
equilibrium with one another is employed.
Fluid/wall friction, gravitational force, and heat
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
4 | P a g e
transfer between the fluid and the outer well
layers are also taken into account.
2. Development of mathematical model
Fig. 1 shows a schematic flow diagram of a
typical injection well. A control volume
consisting of a section of the tube is considered
for the analysis and derivation of model
governing equations.
Fig. 1: Schematic representation of a control
volume within a vertical pipe and the forces
acting on it
Where 𝐹𝑃, 𝐹𝑓, 𝐹𝑔, 𝜌 and 𝑢 are pressure
force, frictional force, gravitational force, fluid
density and velocity respectively. L, Dp and ∆𝑥
are well depth, diameter and differential
control volume.
This study considers a purely vertical
injection tube only hence, pipe inclination is
unaccounted for. The following simplified
assumptions are applied:
One-dimensional flow in the pipe
Homogeneous equilibrium fluid flow
Negligible fluid structure interaction
through vibrations
Constant cross section area of pipe
The assumption of homogeneous
equilibrium flow means that all phases are at
mechanical and thermal equilibrium (i.e.
phases are flowing with same velocity and
temperature) hence the three conservation
equations should be applied for the fluid
mixture. Although, in practice usually the
vapour phase travels faster than the liquid
phase, the HEM model has been investigated
proven to have an acceptable accuracy in many
practical applications.
The following gives a detailed account of
the HEM model employed for the simulation of
the time-dependent flow of CO2 in injection
wells. The system of four partial differential
equations for the CO2 liquid/gas mixture, to be
solved in the well tubing, can be written in the
well-known conservative form as follows:
𝜕
𝜕𝑡𝑸 +
𝜕
𝜕𝑥𝑭(𝑸) = 𝑺𝟏 + 𝑺𝟐 (1)
where
𝑸 = (
𝜌𝐴𝜌𝑢𝐴𝜌𝐸𝐴𝐴
) , 𝑭(𝑸)
= (
𝜌𝑢𝐴
𝜌𝑢2𝐴 + 𝐴𝑃𝜌𝑢𝐻𝐴0
), 𝑺𝟏
=
(
0
𝑃𝜕𝐴
𝜕𝑥00 )
, 𝑺𝟐
= (
0𝐴(𝐹 + 𝜌𝛽𝑔)
𝐴(𝐹𝑢 + 𝜌𝑢𝛽𝑔 + 𝑞)0
)
(2)
In the above, the first three equations
correspond to mass, momentum, and energy
conservation, respectively. The fourth equation
describes the fact that the cross-sectional area
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
5 | P a g e
𝐴 is, at any location along the well, constant in
time, but might vary along the depth of the
well. Moreover, 𝑢 and 𝜌 are the mixture
velocity and density, respectively. 𝑃 is the
mixture pressure, while 𝐸 and 𝐻 represent the
specific total energy and total enthalpy of the
mixture, respectively. They are defined as:
𝐸 = 𝑒 + 1
2 𝑢2 (3)
𝐻 = 𝐸 + 𝑃
𝜌 (4)
where 𝑒 is the specific internal energy. In
addition, 𝑥 denotes the space coordinate, 𝑡 the
time, 𝐹 the viscous friction force, 𝑞 the heat
flux, and 𝑔 the gravitational acceleration. In the
case of the HEM, the assumption of
mechanical equilibrium, i.e. no phase slip, is
retained.
The system of partial differential Eq. (2) is
an extension of the work previously done in
(Curtis M. Oldenburg, 2007), (Celia &
Nordbotten, 2009), by accounting for both a
variable cross-sectional area and additional
source terms. By analysing in more detail the
various source terms appearing on the right-
hand side of Eq. (2).
The frictional loss 𝐹 in Eq. (2) can be
expressed as
𝐹 = −𝑓𝑤𝜌𝑢2
𝐷𝑝 (5)
where 𝑓𝑤 is the Fanning friction factor,
calculated using Chen’s correlation (Chen,
1979), and 𝐷𝑝 is the internal diameter of the
pipe.
The gravitational term includes
𝛽 = 𝜌𝑔 sin 𝜃 (6)
which accounts for the possible well
deviation.
In Eq. (2) the source term 𝑄 accounts for the
heat exchange between the fluid and the well
wall. The corresponding heat transfer
coefficient 𝜂 is calculated using the well-
known Dittus-Boelter correlation (Dittus &
Boelter, 1930):
𝜂 = 0.023 𝑅𝑒0.8𝑃𝑟0.4𝑘
𝐷𝑝 (7)
where 𝑘, 𝑅𝑒 and 𝑃𝑟 are the thermal
conductivity, Reynold’s number and Prandtl’s
number for the fluid. The heat exchanged
between the fluid and the wall is calculated
using the following formula:
𝑞 = 4
𝐷𝑝𝜂(𝑇𝑤 − 𝑇) (8)
𝑇𝑤 and 𝑇 are the temperatures of the fluid
and of the wall, respectively. Note that 𝑇𝑤 =𝑇𝑤(𝑥, 𝑡), i.e. 𝑇𝑤 is not assumed constant, but
variable with time and space.
2.1. Boundary conditions
In order to close the conservation Eq. (2),
suitable boundary conditions must be specified.
In this study, the boundary conditions are
implemented by introducing ’ghost cells’ on
both ends of the computational domain,
representing the wellhead and bottom-hole.
2.1.1. Top of the well
At the wellhead, Fig. 2 is a schematic
representation of the flow through the
wellbore. As can be seen in Fig. 2, CO2 is
injected into the well at the top and exit at the
bottom into the reservoir. The injection strings
are tapered and narrower going down the well.
This allows the rapid injection of CO2 and
minimises the pressure drop along the
wellbore. The development of the required
inlet boundary condition for simulating CO2
injection is detailed below.
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
6 | P a g e
Figure 2: Schematic diagram of the
Goldeneye CO2 injection well (Li et al., 2015)
Injection of high pressure CO2 into a low
pressure well is similar to pipeline
depressurisation due to a puncture or rupture.
As such, the vapour and liquid phases are
dispersed at the well inlet due to high velocity.
Given the above, the HEM assumption is
applied in the ghost cell at the well inlet and
the flow is assumed to be isentropic. As a
result, the conservation equations are rewritten
as (K W Thompson, 1990; Kevin W.
Thompson, 1987):
𝜕𝜌
𝜕𝑡+
1
𝑐2[𝜁2 +
1
2(𝜁3 + 𝜁1)] = 0
(9)
𝜕𝑝
𝜕𝑡+1
2[𝜁3 + 𝜁1] = 0
(10)
𝜕𝑢
𝜕𝑡+
1
2𝜌𝑐[𝜁3 − 𝜁1] − 𝑔 = 0
(11)
Where, 𝜁 is referred to as the wave
amplitude, for which the generic expressions
are given by:
𝜁1 = (𝑢 − 𝑐) {𝜕𝑝
𝜕𝑥− 𝜌𝑐
𝜕𝑢
𝜕𝑥}
(12)
𝜁2 = 0
𝜁3 = (𝑢 + 𝑐) {𝜕𝑝
𝜕𝑥+ 𝜌𝑐
𝜕𝑢
𝜕𝑥}
(13)
𝜁2 describes the inflow entropy, so setting
𝜁2 = 0 states that the inflow entropy is constant
in the 𝑥-direction. Following (K W Thompson,
1990), modifications to Eqs. (12) and (13) may
be required at the flow boundaries. In this case,
𝜁1 and 𝜁3 have to be specified in accordant with
the flow acceleration through the variable flow
areas (i.e. from the injection pipeline into the
wellhead).
From Eqs. (9) and (11), the following can be
derived for 𝜁1 and 𝜁3:
𝜁1 =1
(𝑢−𝑐){2𝜌𝑐2𝑔
𝜕�̇�
𝜕𝑡− (𝑢 + 𝑐)𝜁3}
(14)
𝜁3 =1
(𝑢+𝑐){2𝜌𝑐2𝑔
𝜕�̇�
𝜕𝑡− (𝑢 − 𝑐)𝜁1}
(15)
Where �̇� = 𝜌𝑢 and its derivative with
respect to time is approximated by:
𝜕�̇�
𝜕𝑡=
1
∆𝑡[�̇�|𝑛 − �̇�|𝑛−1]
(16)
where the subscripts, 𝑛 − 1 and 𝑛
respectively denote the previous and current
time-steps.
Finally, Eqs. (9) to (11) can now be readily
solved for a complete description of the flow
conditions in the ghost cell at the wellhead.
2.1.2. Bottom of the well
The boundary condition at the bottom of the
well is described using an empirical pressure-
flow relationship derived from reservoir
properties (Xiaolu Li et al., 2015; Shell, 2015):
�̃� + �̃� × 𝑀 + �̃� × 𝑀2 = 𝑃𝐵𝐻𝑃2 − 𝑃𝑟𝑒𝑠
2 (17)
where
�̃� is the minimum pressure required for the
flow to start from the well into the reservoir, Pa
s/kg (see Table 1)
�̃� and �̃� are site-specific dimensional
constants, Pa.s/kg (see Table 1)
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
7 | P a g e
𝑀 is the instantaneous mass flow rate at the
bottom-hole, kg/s
𝑃𝐵𝐻𝑃 is the instantaneous bottom-hole
pressure, bar and
𝑃𝑟𝑒𝑠 is the reservoir static pressure, bar.
The right hand side of Eq. (17) represent the
pressure differential between the bottom of the
well and the reservoir. Significantly, when
𝑃𝐵𝐻𝑃2 > 𝑃𝑟𝑒𝑠
2 there is injection into the
reservoir whereas when 𝑃𝐵𝐻𝑃2 < 𝑃𝑟𝑒𝑠
2 there is a
backflow or blowout. In other words, the
boundary condition at the bottom of the well
can be discretely utilised to satisfy both
injection and blowout cases. In describing the
relationship between the reservoir and the
bottom of the well, Eq. (17) represents a more
sophisticated condition than a standard, linear
correlation between the bottom-hole pressure
and the flow rate.
3. Numerical method and injection well CO2
inlet conditions
In this study, an effective model based on
the Finite Volume Method (FVM),
incorporating a conservative Godunov type
finite-difference scheme (Godunov 1959,
Radvogin et al. 2011, Cumber et al. 1994) is
used. The FVM is well-established and
thoroughly validated CFD technique. In
essence, the methodology involves the
integration of the fluid flow equations over the
entire control volumes of the solution domain
and then accurate calculation of the fluxes
through the boundaries of the computed cells.
For the purpose of numerical solution of the
governing equations they are written in a
vector form (Toro 2010):
𝜕�⃗�
𝜕𝑡+𝜕𝑓
𝜕𝑧= 𝑆 ,
(18)
where
�⃗� = (𝜌 , 𝜌𝑢 , 𝜌𝑒)𝑇,
𝐹 = (𝜌𝑢 , (𝜌𝑢2 + 𝑃) ,
𝑢( 𝜌𝑢𝑒 + 𝜌𝑢2 + 𝑃))𝑇
𝑆 = (𝑆𝑚, 𝑆𝑚𝑜𝑚, 𝑆𝑒)𝑇
(19)
�⃗� , 𝐹 and 𝑆 are the vectors of conserved
variables, fluxes and source terms respectively.
The source terms 𝑆𝑚, 𝑆𝑚𝑜𝑚and 𝑆𝑒 describe the
effects of mass, momentum and energy
exchange between the fluid and its surrounding
respectively, as well as friction and heat
exchange at the pipe wall.
3.1 Model Validation
The model relevance and applicability to
real-world injection project is validated using
Ketzin pilot site experimental data. The model
predictions are closely in agreement with the
real-life CO2 injection scenario.
CO2 injection well initial and boundary
conditions for Ketzin pilot site Brandenburg,
Germany obtained from Möller et al, (2014)
employed for the model validation are:
CO2 inlet pressure 57 bar, temperatures
10 oC and 20 oC, and injection mass
flow rate 0.41 kg/s
Initial wellhead pressure 48 bar and
temperature 10 oC,
Total well depth 550m; 0.0889m
internal diameter
Initial bottom-hole pressure 68 bar and
temperature 33 oC,
As can be seen in Figs. 3 and 4, the
simulation results for 10oC and 20oC injection
inlet temperature condition all showed good
agreement with the experimental data. The
performance of our model in relation to the
Ketzin pilot real-world experimental injection
project shows it reliability and applicability.
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
8 | P a g e
Figure 3: Well temperature profile for 10oC
inlet CO2 injection temperature
Figure 4: Well temperature profile for 20oC
inlet CO2 injection temperature
4. Injection well and CO2 inlet conditions
The data used in this study obtained from
the Peterhead CCS project include the well
depth and pressure and temperature profiles,
along with the surrounding formation
characteristics as presented in (Xiaolu Li et al.,
2015; Shell UK, 2015) and reproduced in Table
1. The Peterhead CCS project was proposed to
capture one million tonnes of CO2 per annum
for 15 years from an existing combined cycle
gas turbine located at Peterhead Power Station
in Aberdeenshire, Scotland. In the project, the
CO2 captured from the Peterhead Power
Station would have been transported by
pipeline and then injected into the depleted
Goldeneye reservoirs. Despite the cancellation
of the project funding, useful information was
already available, given that the Goldeneye
reservoir had been used for extraction of
natural gas for many years.
Table 1: Goldeneye injection well and CO2
inlet conditions (Shell UK, 2015)
Input parameter Value
Wellhead pressure,
bar 36
Wellhead
temperature, K 277
Bottom-hole
temperature, K 353
Well
diameter/depth, m
0 – 800m depth:
0.125m
800 – 2000m depth:
0.0765m
2000–2582m depth:
0.0625m
CO2 injection mass
flow rate, kg/s
3 different cases:
linearly ramped-up
to 38.5 kg/s in 5
minutes, 30 minutes
and 2 hours
Injection tube
diameter, m 0.125m
CO2 inlet pressure,
bar 115
CO2 inlet
temperature, K 280
Outflow
�̃� = 0
�̃� = 1.3478 × 1012 Pa2s/kg
�̃� = 2.1592 × 1010 Pa2s2/kg2
𝑃𝑟𝑒𝑠 = 172 bar
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
9 | P a g e
Fig. 5 shows the initial distribution of the
pressure and the temperature along the well.
Based on this initial pressure and temperature
distribution obtained from the Goldeneye CO2
injection well, the CO2 is in the gas phase in
the first 400 m along the well and dense phase
within the rest of the well depth. This initial
condition is the same for all 3 ramp-up
injection cases considered. In particular, this
study considers linearly ramped-up injection
mass flow rates from 0 to 38.5 kg/s in 5
minutes, 30 minutes and 2 hours. The varying
ranges of CO2 injection mass flow rate ramped
up with time from 0 to 38.5 kg/s at CO2 feed
pressure of 115 bar employed in this study is
essential to understanding the optimum
injection ramp-up duration.
Figure 5: Initial distribution of the pressure
and the temperature along the well
4.1 Thermodynamics
In this study, the CO2 thermal properties and
phase equilibrium data are computed using
Reng-Robinson (PR) Equation of State (EoS)
(Peng & Robinson, 1976). PR EoS has been
extensively applied and validated in pipeline
decompression studies especially for CO2
showing very good agreement with
experimental data (see Brown et al, 2013a;
Brown et al, 2013b; Mahgerefteh et al, 2012;
Martynov et al, Sundara, 2013). The PR EoS
has been proven valid and highly useful for
both single fluid and multi-component studies
hence the validity of this study (see also Hajiw
et al, 2018; Henderson et al, 2001; IEAGHG,
2011; Kwak & Kim, 2017; Mireault et al,
2010; Wang et al, 2011; Wang et al, 2015;
Ziabakhsh-Ganji & Kooi, 2014).
5. Results and discussion
The results obtained following the
simulation of the transient flow model for the
ramp-up injection of CO2 into highly depleted
oil/gas fields are presented and discussed in
details. Beginning at the top inlet of the
injection well down to the bottom outlet into
the reservoir, the pressure and temperature
profiles are presented. Specifically, the
pressure and temperature variations at the top
of the well and the corresponding effects along
the wellbore can be seen in the profiles.
Figs. 6 to 8 show the variations in pressure
with time at the top of the well for 5 mins, 30
mins and 2 hrs ramp-up injection times. The
start-up injection data shows a rapid drop in
pressure of the incoming CO2 at 115 bar to as
low as 36.5 bar. The pressure drops to 74, 66
and 36.5 bar corresponding to injection ramp-
up period of 5 mins, 30 mins and 2hrs
respectively. Following the initial pressure is a
recovery over the ramp-up duration until a
steady state is attained in each case. The
pressure recovery is quicker for the fast 5 mins
ramp-up case and slower for the 30 mins or 2
hrs ramp-up cases. Consequently a fast ramp-
up during CO2 injection will help minimise the
initial pressure drop due to the difference in
wellhead pressure and that of incoming CO2
and also enhance a fast recovery after rapid
pressure drop.
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
10 | P a g e
Figure 6: CO2 wellhead pressure variation with
time for 5 mins ramp-up injection case
Figure 7: CO2 wellhead pressure variation with
time for 30 mins ramp-up injection case
Figure 8: CO2 wellhead pressure variation
with time for 2 hrs ramp-up injection case
As can be seen in Fig. 9 where all three
ramp-up cases are shown together. The 5 mins
ramp-up case predicts a lesser pressure drop
and high recovery compared with the other
cases of 30 mins and 2 hrs ramp-up. This
implies that for best practice a fast ramp-up is
recommended to minimise the risk of large
pressure drop and a corresponding low
temperature. This shows that the injected CO2
undergoes a refrigeration effect caused by the
Joule-Thomson expansion.
Figure 9: CO2 wellhead pressure profiles for 5
mins, 30 mins and 2 hrs ramp-up injection
cases
Figs. 10 to 12 show the variations in
temperature with time at the top of the well for
5 mins, 30 mins and 2 hrs ramp-up injection
times. The start-up injection data shows a rapid
drop in temperature of the incoming CO2 from
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
11 | P a g e
277 K to as low 226 K. The temperature drops
to 252, 238 and 226 K corresponding to
injection ramp-up period of 5 mins, 30 mins
and 2hrs respectively. Following the initial
temperature drop is a recovery over the ramp-
up duration until a steady state is attained in
each case. The temperature recovery however
gets better with increasing ramp-up duration.
The predicted steady-state temperatures after
recovery for the fast 5 mins ramp-up case and
slower for the 30 mins or 2 hrs ramp-up cases
are respectively 277, 280 and 303 K.
Consequently a fast ramp-up during CO2 start-
up injection will help minimise this initial
temperature drop due to the difference in the
pressure at top of the injection well and that of
the incoming CO2. Also, in order to enhance a
fast recovery after the initial rapid pressure and
temperature drop.
Figure 10: CO2 wellhead temperature variation
with time for 5 mins ramp-up injection case
Figure 11: CO2 wellhead temperature variation
with time for 30 mins ramp-up injection case
Figure 12: CO2 wellhead temperature variation
with time for 2 hrs ramp-up injection case
As can be seen in Fig. 13 where all three
ramp-up cases of CO2 wellhead temperature
profiles are shown together. The 5 mins ramp-
up case predicts a lesser temperature drop
compared with the other cases of 30 mins and 2
hrs ramp-up. This implies that for best practice
a fast (5 mins) ramp-up is recommended to
minimise the risk of large temperature drop. In
simply terms, the fast ramp-up is exposed to a
short time refrigeration effect caused by the
Joule-Thomson expansion.
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
12 | P a g e
Figure 13: CO2 wellhead temperature profiles
for 5 mins, 30 mins and 2hrs ramp-up injection
cases
Figs. 14 and 15 respectively show the
corresponding results for the temperature and
pressure profiles along the length of the well
during the 5 mins ramping up process at
different selected time intervals of 10, 100, 200
and 300 seconds. The pressure profiles show
continues pressure build up along the wellbore
during the 5 mins injection ramp up. On the
contrary, the temperature profiles show a
significant temperature drop for the 10, 100
and 200 s well profiles in comparison with the
initial well temperature profile at 0 s. The well
temperature profile at 300 s shows significant
recovery approaching the initial well
temperature profile. As can been seen in Fig.
15 the temperature profile along the injection
wellbore shows an improve profile as the ramp
up injection duration reaches 300 s compared
with the 10, 100 and 200 s cases. In other
words, operating a fast injection ramp-up is
recommended to rapidly increase the injection
flow rate with time and consequently minimise
the drop in temperature along the wellbore.
Figure 14: 5 mins ramp-up injection well
pressure profiles at varying times
Figure 15: 5 mins ramp-up injection well
temperature profiles at varying times
Finally, the current model is briefly compared
with OLGA simulation result that Acevedo &
Chopra, (2017) presented in their paper
“Influence of phase behaviour in the well
design of CO2 injectors”.
In Fig. 16, the results of both model simulation
indicated low temperature in the top section of
the well during the continuous injection period.
The reduction in temperature is due to the
flashing of the liquid CO2 to gas/liquid CO2
caused by the low reservoir pressure. During
transient operations (closing-in and re-starting
injection operations), a temperature drop is
observed at the top of the well for a short
period of time. The sequence of steady state
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
13 | P a g e
injection, closing-in operation (30 minutes),
closed-in time (10minutes), and starting-up
operation (30minutes) for the low reservoir
pressure case was simulated using OLGA
(Acevedo & Chopra, 2017) and our model, see
Figure 16. As can be seen the current model
predicted slightly lower CO2 temperatures than
OLGA. This is likely due to the inconsistency
of the Peng-Robinson Equation of State
employed in the current study for the
prediction of CO2 thermodynamics properties.
Figure 16: Comparison of temperature profiles
at the top of the well sequence of closing and
opening a CO2 injector well
6. Conclusions and recommendations
This study has led to the development and
testing of a rigorous model for the simulation
of the highly-transient multi-phase flow
phenomena taking place in wellbores during
the start-up injection of high pressure CO2 into
depleted gas fields. In practice, the model
developed can serve as a valuable tool for the
development of optimal injection strategies and
best-practice guidelines for the minimisation of
the risks associated with the start-up injection
of CO2 into depleted gas fields. In order to
assess key parameters having the greatest
impact on the transient flow of CO2 in the
injection well, sensitivity analysis has been
performed over several parameters by different
authors. However, varying injection ramping
up times considered in this study has never
been studied in previous publications. As such,
this study introduces a new area of
consideration and gives a clearer insight on the
effects of CO2 injection ramp-up times on the
wellbore pressure and temperature profiles.
The importance of investigating the injection
flow rates ramp-up times is driven by the need
for the development of optimal injection
strategies and best-practice guidelines for the
minimisation of the risks associated with the
process.
Based on the application of the model
developed in this work to realistic test cases
involving the ramping up of CO2 injection flow
rates from 0 to 38.5 kg/s into the Goldeneye
depleted reservoir in the North Sea following
fast (5 mins), medium (30 mins) and slow (2
hrs) injection ramp-up times, the main findings
and recommendations of this work can be
summarised as follows:
The degree of cooling along the
injection well becomes less severe with
a decrease in the injection ramp-up
duration. In other words, operating a
fast start-up injection ramp-up is
recommended (i.e. between 5 and 30
mins) rather than a slower one (over 2
hrs). This is due the limited exposure
time to the refrigeration effect caused
by the JT expansion on the injected
CO2.
The formation of ice during the ramp
up injection process is likely, given that
in all cases considered the minimum
fluid temperature falls well below 0 oC
at the top of the well. This poses the
risk of ice formation and ultimately
well blockage should a sufficient
quantity of water be present. Thus, for
this case study injecting CO2 at
temperatures well above 290 K at flow
rate ramp-up duration of 5 mins will be
best practice to minimise this risk.
The minimum simulated CO2
temperature and the corresponding
pressure predicted at the top of well
during the start-up injection flow rate
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
14 | P a g e
ramp-up process are close to the ranges
where CO2 hydrates formation would
be expected. The ideal hydrate
formation conditions are pressures and
temperatures below 273.15 K and 12.56
bar however, tiny molecules of CO2
hydrates begin formation at pressures
and temperatures just below 283 K and
49.99 bar (Circone et al., 2003; Yang et
al., 2012).
Given the observed relatively large
drop in temperature at the top of the
well especially for the 2 hrs injection
flow ramp-up case, well failure due to
thermal stress shocking (i.e. as a result
of the temperature gradient between
inner and outer layers of the steel
casing) during the injection process
could be a real risk. This could lead to a
brittle fracture situation should the
temperature drops below –30 oC. The
design temperature at which steel
becomes highly brittle according the
Stainless Steel Information Centre of
North America. Hence, the risk of
injection system failure due to thermal
stress shocking of the steel casing can
be avoided by keeping to a fast or
medium injection flow ramp-up.
Finally, it is critically important to bear in mind
that the above conclusions are not universal.
On the contrary, they are only based on the
case study investigated. Each injection scenario
must be individually examined in order to
determine the likely risks. In this paper, the
necessary computational tool is developed to
make such risks assessment.
7. References
Acevedo, L., & Chopra, A. (2017). Influence
of Phase Behaviour in the Well Design of
CO2Injectors. Energy Procedia,
114(November 2016), 5083–5099.
http://doi.org/10.1016/j.egypro.2017.03.16
63
André, L., Audigane, P., Azaroual, M., &
Menjoz, A. (2007). Numerical modeling
of fluid-rock chemical interactions at the
supercritical CO2-liquid interface during
CO2 injection into a carbonate reservoir,
the Dogger aquifer (Paris Basin, France).
Energy Conversion and Management,
48(6), 1782–1797.
http://doi.org/10.1016/j.enconman.2007.0
1.006
Böser, W., & Belfroid, S. (2013). Flow
Assurance Study. Energy Procedia, 37,
3018–3030.
http://doi.org/10.1016/j.egypro.2013.06.18
8
Brown, S., Beck, J., Mahgerefteh, H., & Fraga,
E. S. (2013). Global sensitivity analysis of
the impact of impurities on CO2 pipeline
failure. Reliability Engineering and
System Safety, 115, 43–54.
http://doi.org/10.1016/j.ress.2013.02.006
Brown, S., Martynov, S., Mahgerefteh, H., &
Proust, C. (2013). A homogeneous
relaxation flow model for the full bore
rupture of dense phase CO2 pipelines.
International Journal of Greenhouse Gas
Control, 17, 349–356.
http://doi.org/10.1016/j.ijggc.2013.05.020
Celia, M. a., & Nordbotten, J. M. (2009).
Practical modeling approaches for
geological storage of carbon dioxide.
Ground Water, 47(5), 627–638.
http://doi.org/10.1111/j.1745-
6584.2009.00590.x
Chen, N. H. (1979). An Explicit Equation for
Friction Factor in Pipe. Industrial &
Engineering Chemistry Fundamentals,
18(3), 296–297.
http://doi.org/10.1021/i160071a019
Circone, S., Stern, L. a., Kirby, S. H., Durham,
W. B., Chakoumakos, B. C., Rawn, C. J.,
… Ishii, Y. (2003). CO 2
Hydrate: Synthesis, Composition,
Structure, Dissociation Behavior, and a
Comparison to Structure I CH 4 Hydrate.
The Journal of Physical Chemistry B,
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
15 | P a g e
107(23), 5529–5539.
http://doi.org/10.1021/jp027391j
COP 21. (2015). COP 21 Paris France
Sustainable Innovation Forum 2015
working with UNEP.
Cotton, A., Gray, L., & Maas, W. (2017).
Learnings from the Shell Peterhead CCS
Project Front End Engineering Design.
Energy Procedia, 114, 5663–5670.
http://doi.org/10.1016/J.EGYPRO.2017.0
3.1705
De Koeijer, G., Hammer, M., Drescher, M., &
Held, R. (2014). Need for experiments on
shut-ins and depressurizations in
CO2injection wells. In Energy Procedia
(Vol. 63, pp. 3022–3029). Elsevier B.V.
http://doi.org/10.1016/j.egypro.2014.11.32
5
Dennis Gammer. (2016). Reducing the cost of
CCS - Developments in Capture Plant… |
The ETI. Retrieved from
http://www.eti.co.uk/insights/reducing-
the-cost-of-ccs-developments-in-capture-
plant-technology/
Dittus, F. W., & Boelter, L. M. K. (1930). Heat
Transfer in Automobile Radiators, 2, 443.
Goodarzi, S., Settari, A., Zoback, M. D., &
Keith, D. (2010). Thermal Aspects of
Geomechanics and Induced Fracturing in
CO2 Injection With Application to CO2
Sequestration in Ohio River Valley. SPE
International Conference on CO2
Capture, Storage, and Utilization. Society
of Petroleum Engineers.
http://doi.org/10.2118/139706-MS
Hajiw, M., Corvisier, J., Ahmar, E. El, &
Coquelet, C. (2018). Impact of impurities
on CO2 storage in saline aquifers:
Modelling of gases solubility in water.
International Journal of Greenhouse Gas
Control, 68, 247–255.
http://doi.org/10.1016/J.IJGGC.2017.11.0
17
Hannis, S., Pearce, J., Chadwick, A., & Kirk,
K. (2017). IEAGHG Technical Report
January 2017 Case Studies of CO 2
Storage in Depleted Oil and Gas Fields,
(January).
Henderson, N., de Oliveira, J. R., Souto, H. P.
A., & Marques, R. P. (2001). Modeling
and Analysis of the Isothermal Flash
Problem and Its Calculation with the
Simulated Annealing Algorithm.
Industrial & Engineering Chemistry
Research, 40(25), 6028–6038.
http://doi.org/10.1021/ie001151d
Hughes, D. S. (2009). Carbon storage in
depleted gas fields: Key challenges.
Energy Procedia, 1(1), 3007–3014.
http://doi.org/10.1016/j.egypro.2009.02.07
8
IEAGHG. (2015). Carbon Capture and Storage
Cluster Projects: Review and Future
Opportunities, (April).
IEAGHG (International Energy Agency
greenhouse gas emissions). (2011).
Effects of Impurities on Geological
Storage of CO2, (June), Report: 2011/04.
Jiang, P., Li, X., Xu, R., Wang, Y., Chen, M.,
Wang, H., & Ruan, B. (2014). Thermal
modeling of CO2 in the injection well and
reservoir at the Ordos CCS demonstration
project, China. International Journal of
Greenhouse Gas Control, 23, 135–146.
http://doi.org/10.1016/j.ijggc.2014.01.011
Kwak, D.-H., & Kim, J.-K. (2017). Techno-
economic evaluation of CO2 enhanced oil
recovery (EOR) with the optimization of
CO2 supply. International Journal of
Greenhouse Gas Control, 58, 169–184.
http://doi.org/10.1016/j.ijggc.2017.01.002
Li, X., Li, G., Wang, H., Tian, S., Song, X., Lu,
P., & Wang, M. (2017). International
Journal of Greenhouse Gas Control A
unified model for wellbore flow and heat
transfer in pure CO 2 injection for
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
16 | P a g e
geological sequestration , EOR and
fracturing operations. International
Journal of Greenhouse Gas Control, 57,
102–115.
http://doi.org/10.1016/j.ijggc.2016.11.030
Li, X., Xu, R., Wei, L., & Jiang, P. (2015).
Modeling of wellbore dynamics of a CO2
injector during transient well shut-in and
start-up operations. International Journal
of Greenhouse Gas Control, 42, 602–614.
http://doi.org/10.1016/j.ijggc.2015.09.016
Lindeberg, E. (2011). Modelling pressure and
temperature profile in a CO2 injection
well. Energy Procedia, 4, 3935–3941.
http://doi.org/10.1016/j.egypro.2011.02.33
2
Linga, G., & Lund, H. (2016). A two-fluid
model for vertical flow applied to CO2
injection wells. International Journal of
Greenhouse Gas Control, 51, 71–80.
http://doi.org/10.1016/j.ijggc.2016.05.009
Lu, M., & Connell, L. D. (2008). Non-
isothermal flow of carbon dioxide in
injection wells during geological storage.
International Journal of Greenhouse Gas
Control, 2(2), 248–258.
http://doi.org/10.1016/S1750-
5836(07)00114-4
Lu, M., & Connell, L. D. (2014). The transient
behaviour of CO2 flow with phase
transition in injection wells during
geological storage – Application to a case
study. Journal of Petroleum Science and
Engineering, 124, 7–18.
http://doi.org/10.1016/j.petrol.2014.09.02
4
Mahgerefteh, H., Brown, S., & Denton, G.
(2012). Modelling the impact of stream
impurities on ductile fractures in CO2
pipelines. Chemical Engineering Science,
74, 200–210.
http://doi.org/10.1016/j.ces.2012.02.037
Martynov, S., Brown, S., Mahgerefteh, H., &
Sundara, V. (2013). Modelling choked
flow for CO2 from the dense phase to
below the triple point. International
Journal of Greenhouse Gas Control, 19,
552–558.
http://doi.org/10.1016/j.ijggc.2013.10.005
Mireault, R. A., Stocker, R., Dunn, D. W., &
Pooladi-Darvish, M. (2010). Wellbore
Dynamics of Acid Gas Injection Well
Operation. Society of Petroleum
Engineers. http://doi.org/10.2118/135455-
MS
Möller, F., Liebscher, A., Martens, S.,
Schmidt-Hattenberger, C., & Streibel, M.
(2014). Injection of CO2 at ambient
temperature conditions - Pressure and
temperature results of the “cold injection”
experiment at the Ketzin pilot site. In
Energy Procedia (Vol. 63, pp. 6289–
6297).
http://doi.org/10.1016/j.egypro.2014.11.66
0
Munkejord, S. T., Hammer, M., & L??vseth, S.
W. (2016). CO2 transport: Data and
models - A review. Applied Energy, 169,
499–523.
http://doi.org/10.1016/j.apenergy.2016.01.
100
Nordbotten, J. M., Celia, M. A., & Bachu, S.
(2005). Injection and storage of CO2 in
deep saline aquifers: Analytical solution
for CO2 plume evolution during injection.
Transport in Porous Media, 58(3), 339–
360. http://doi.org/10.1007/s11242-004-
0670-9
Oldenburg, C. M. (2007). Joule-Thomson
cooling due to CO2 injection into natural
gas reservoirs. Energy Conversion and
Management, 48(6), 1808–1815.
http://doi.org/10.1016/j.enconman.2007.0
1.010
Oldenburg, C. M., Pruess, K., & Benson, S. M.
(2001). Process modeling of CO2
Samuel and Mahgerefeth / International Journal of Greenhouse Gas Control
17 | P a g e
injection into natural gas reservoirs for
carbon sequestration and enhanced gas
recovery. Energy & Fuels, 15(2), 293–
298. http://doi.org/Doi
10.1021/Ef000247h
Pan, L., Webb, S. W., & Oldenburg, C. M.
(2011). Analytical solution for two-phase
flow in a wellbore using the drift-flux
model. Advances in Water Resources,
34(12), 1656–1665.
http://doi.org/10.1016/j.advwatres.2011.0
8.009
Paterson, L., Lu, M., Connell, L., & Ennis-
King, J. (2008). Numerical Modeling of
Pressure and Temperature Profiles
Including Phase Transitions in Carbon
Dioxide Wells. SPE Annual Technical
Conference and Exhibition, Denver,
Colorado USA, Sept 21-24. Society of
Petroleum Engineers.
http://doi.org/10.2118/115946-MS
Peng, D.-Y., & Robinson, D. B. (1976). A New
Two-Constant Equation of State.
Industrial & Engineering Chemistry
Fundamentals, 15(1), 59–64.
http://doi.org/10.1021/i160057a011
Ruan, B., Xu, R., Wei, L., Ouyang, X., Luo, F.,
& Jiang, P. (2013). Flow and thermal
modeling of CO2 in injection well during
geological sequestration. International
Journal of Greenhouse Gas Control, 19,
271–280.
http://doi.org/10.1016/j.ijggc.2013.09.006
Samuel, R. J., & Mahgerefteh, H. (2017).
Transient Flow Modelling of Start-up
CO2 Injection into Highly-Depleted
Oil/Gas Fields. International Journal of
Chemical Engineering and Applications,
8(5), 319–326.
http://doi.org/10.18178/ijcea.2017.8.5.677
Sanchez Fernandez, E., Naylor, M., Lucquiaud,
M., Wetenhall, B., Aghajani, H., Race, J.,
& Chalmers, H. (2016). Impacts of
geological store uncertainties on the
design and operation of flexible CCS
offshore pipeline infrastructure.
International Journal of Greenhouse Gas
Control, 52, 139–154.
http://doi.org/10.1016/j.ijggc.2016.06.005
Shell. (2015). Peterhead CCS Project.
Shell UK, T. R. (2015). Peterhead CCS Project.
Thompson, K. W. (1987). Time dependent
boundary conditions for hyperbolic
systems. Journal of Computational
Physics, 68(1), 1–24.
http://doi.org/10.1016/0021-
9991(87)90041-6
Thompson, K. W. (1990). Time-dependent
boundary conditions for hyperbolic
systems, {II}. Journal of Computational
Physics, 89, 439–461.
http://doi.org/10.1016/0021-
9991(90)90152-Q
UNFCCC. (2016). The Paris Agreement - main
page.
Wang, J., Ryan, D., Anthony, E. J., Wildgust,
N., & Aiken, T. (2011). Effects of
impurities on CO2transport, injection and
storage. Energy Procedia, 4, 3071–3078.
http://doi.org/10.1016/j.egypro.2011.02.21
9
Wang, J., Wang, Z., Ryan, D., & Lan, C.
(2015). A study of the effect of impurities
on CO2 storage capacity in geological
formations. International Journal of
Greenhouse Gas Control, 42, 132–137.
http://doi.org/10.1016/J.IJGGC.2015.08.0
02
Yang, M., Song, Y., Ruan, X., Liu, Y., Zhao,
J., & Li, Q. (2012). Characteristics of CO2
hydrate formation and dissociation in
glass beads and silica gel. Energies, 5(4),
925–937.
http://doi.org/10.3390/en5040925
Recommended